Systems and methods include steps of determining a state of a network based on telemetry data; determining a value of a reward associated with the state; determining an action to take on the network to bring the network to a next state that is expected to have a better than or equal to value of the reward; and causing the action to be implemented in the network. The steps can also include continuing the determining steps and the causing step.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The non-transitory computer-readable medium of claim 1, wherein the determining the state and the updated state of the network are one or both also performed on demand.
A system and method for network state monitoring and analysis involves determining the current state and an updated state of a network, where these determinations can be performed either periodically or on demand. The network state includes information about network devices, connections, and performance metrics. The system collects data from network devices, analyzes the data to identify changes in network conditions, and generates reports or alerts based on the analysis. The on-demand functionality allows users to trigger state determinations at specific times or in response to particular events, providing flexibility in network monitoring. The system may also include features for visualizing network state changes over time, identifying trends, and predicting potential issues. The technology is designed to improve network management by providing real-time and historical insights into network performance and behavior.
3. The non-transitory computer-readable medium of claim 1, wherein the network is a packet network, and wherein the reward relates to any of throughput, latency, jitter, workload, dropped packets, and packet errors.
This invention relates to a system for optimizing network performance in packet networks by dynamically adjusting network parameters based on performance metrics. The system monitors key performance indicators such as throughput, latency, jitter, workload, dropped packets, and packet errors to assess network conditions. Based on these metrics, the system calculates a reward value that quantifies the effectiveness of current network configurations. The reward value is then used to adjust network parameters, such as routing paths, bandwidth allocation, or quality of service settings, to improve overall network performance. The system may employ machine learning or reinforcement learning techniques to refine adjustments over time, ensuring continuous optimization. The goal is to enhance network efficiency, reduce latency, minimize packet loss, and maintain stable performance under varying workloads. The invention is particularly useful in environments where network conditions fluctuate, such as in cloud computing, data centers, or telecommunications networks. By dynamically responding to real-time performance data, the system ensures optimal resource utilization and reliability.
4. The non-transitory computer-readable medium of claim 1, wherein the action includes any of bandwidth changes, re-routing services, and hardware reassignment.
This invention relates to network management systems that dynamically adjust network resources in response to detected conditions. The problem addressed is the need for automated, real-time adjustments to optimize network performance, reduce latency, and improve resource utilization without manual intervention. The system includes a monitoring component that continuously tracks network metrics such as traffic load, latency, and hardware status. When predefined thresholds are exceeded, the system triggers automated actions to mitigate issues. These actions include modifying bandwidth allocations, re-routing network traffic to alternative paths, or reassigning hardware resources to balance the load. The system also supports predictive adjustments based on historical data and machine learning models to anticipate and prevent potential bottlenecks. The invention ensures efficient resource allocation by dynamically adapting to changing network conditions, reducing downtime, and improving overall system reliability. It is particularly useful in large-scale networks where manual adjustments are impractical or inefficient. The automated nature of the system allows for rapid responses to network fluctuations, enhancing performance and user experience.
11. The method of claim 10, wherein the determining the state and the updated state of the network are one or both also performed on demand.
A system and method for network state monitoring and analysis involves dynamically determining the state of a network and updating that state based on real-time or on-demand evaluations. The network state includes operational parameters such as connectivity, performance metrics, and resource utilization across nodes and links. The method collects data from network devices, processes the data to identify current conditions, and compares these conditions against predefined thresholds or historical baselines to detect anomalies or inefficiencies. The updated state is then used to optimize network performance, predict failures, or trigger automated corrective actions. The determination of the network state and its updates can be performed continuously in real-time or triggered on demand, allowing for flexible monitoring and adaptive responses to changing network conditions. This approach enhances network reliability, reduces downtime, and improves overall efficiency by providing timely insights into network health and performance. The system may integrate with existing network management tools or operate as a standalone solution, depending on implementation requirements.
12. The method of claim 10, wherein the network is a packet network, and wherein the reward relates to any of throughput, latency, jitter, workload, dropped packets, and packet errors.
This invention relates to optimizing network performance in packet networks by dynamically adjusting network parameters based on performance metrics. The method involves monitoring key performance indicators (KPIs) such as throughput, latency, jitter, workload, dropped packets, and packet errors to assess network efficiency. A reward-based system evaluates these metrics to determine optimal network configurations, such as routing paths, bandwidth allocation, or quality of service (QoS) settings. The system dynamically adjusts these parameters to maximize performance, ensuring efficient data transmission while minimizing disruptions. By continuously analyzing real-time data, the method adapts to changing network conditions, improving reliability and user experience. The approach is particularly useful in environments where network traffic varies significantly, such as in cloud computing, telecommunication networks, or enterprise systems. The invention aims to enhance network resilience and efficiency by leveraging performance-based adjustments, reducing manual intervention and improving overall system stability.
13. The method of claim 10, wherein the action includes any of bandwidth changes, re-routing services, and hardware reassignment.
This invention relates to dynamic network management systems that optimize resource allocation in response to changing conditions. The problem addressed is the inefficiency of static network configurations, which fail to adapt to varying demand, leading to underutilized or overloaded resources. The invention provides a method for automatically adjusting network operations based on real-time data to improve performance and reliability. The method involves monitoring network conditions, such as traffic patterns, hardware status, and service demands, to detect deviations from optimal performance. When a deviation is identified, the system triggers an action to rebalance resources. These actions include modifying bandwidth allocation to prioritize critical services, rerouting traffic to avoid congestion, or reassigning hardware components to better match workload requirements. The system may also predict future demand based on historical data and preemptively adjust resources to prevent disruptions. The invention ensures that network resources are used efficiently, reducing costs and improving service quality. By automating adjustments, it minimizes manual intervention, allowing for faster responses to network changes. The solution is applicable to various network environments, including data centers, telecommunications, and cloud computing infrastructures. The dynamic adjustments help maintain high availability and performance, even under fluctuating loads.
17. The method of claim 10, wherein the reward is minimizing dropped packets or packet errors, and wherein the action includes any of i) increasing or decreasing bandwidth of one or more services, ii) re-routing the one or more services to less congested paths, and iii) no action.
This invention relates to optimizing network performance by dynamically adjusting network parameters to minimize packet loss or errors. The method involves monitoring network conditions and applying actions to improve reliability. The reward function is defined as reducing dropped packets or packet errors, ensuring efficient data transmission. Actions include adjusting bandwidth allocation for one or more services, either increasing or decreasing it based on demand and congestion. Another action involves rerouting services to less congested network paths to avoid bottlenecks. Additionally, the system may determine that no action is needed if the network is operating optimally. The method dynamically selects the most effective action to maintain or improve network performance, adapting to real-time conditions. This approach helps prevent data loss and errors, enhancing overall network reliability and user experience. The system continuously evaluates network metrics to decide whether to modify bandwidth, reroute traffic, or take no action, ensuring efficient resource utilization.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
September 9, 2021
April 16, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.